H. vulgaris (strain Zurich) was asexually cultured in Hydra medium (HM: 1 mM CaCl₂ and 0.1 mM NaHCO₃) at pH 7. The animals were kept at 18 ± 1°C and fed thrice a week with freshly hatched Artemia salina nauplii.

For EV collection, 250 polyps were placed in 10 ml of HM in 6 cm Petri dishes. After 4 days starvation, Hydra conditioned medium was collected and centrifuged using a Beckman Coulter Avanti J-25 centrifuge (rotor JA-25.15) at 10,000 g for 30 min to remove tissue patches and at 20,000 g for 30 min to pellet any cellular debris. The resulting supernatant was then centrifuged at 74,200 g for 150 min at 4°C and the pelleted EVs were washed in HM followed by one last centrifugation. Pelleted EVs were resuspended in either PBS or Hydra medium.

EVs were visualized using TEM following a slightly modified protocol to that reported by Thery et al. (2006). Briefly, EVs were fixed with an equal volume of 4% paraformaldehyde and deposited onto a Formvar-carbon coated electron microscopy grid. Once dried, the grids were washed with PBS, fixed with 1% glutaraldehyde for 5 min and washed with distilled water 8 times. The samples were stained with 1% ammonium molybdate for 2 min, dried and observed using a FEI Tecnai 12,120 kV or JEOL United States, Inc., United States) at 80 kV. Hydra polyps were dissected into appropriate pieces to fit into cup-shaped HPF specimen carriers. Tissue pieces were pipetted into a 0.2- or 0.3-mm deep carrier and covered with an additional carrier. Finally, the obtained sandwiches were cryo-immobilized by HPF. Frozen sandwiches were transferred into appropriate containers for storage in Liquid N₂ for later use or subjected to freeze substitution (FS). The frozen samples were transferred under liquid N₂ into cryo-vials containing frozen FS cocktails (anhydrous acetone plus 1% OsO₄ and 0.1–0.2% uranyl acetate). Subsequently, the lids were screwed loosely onto the vials to permit safe evaporation of excess N₂ gas. The vials were placed into the precooled FS device and after about 1 h the lids tightened and FS was performed for at least 8 h at −80 to −90°C; warming up to −55°C at a rate of 5–10°C per hour, subsequent post-fixation and staining at −55°C for 6 h, followed by warming up to −30°C at a rate of 5–10°C per hour where samples were left for an additional 3 h. Finally, samples were washed three times with acetone (10 min each), 16 h in 10% epoxy resin in acetone, 6 h in 30% epoxy resin in acetone, 16 h 70% epoxy resin in acetone, 6 h 100% epoxy resin in acetone, 16 h 100% epoxy resin in acetone. Then the samples were placed in fresh resin and placed in oven at 60°C overnight.

Sample serial thin (70 nm) section of animals were cut with a diamond knife and mounted onto 150–200 mesh Hex grids. Ultrathin sections were examined with transmission EM Libra 120 EFTEM (Zeiss, Oberkochen, Germany) at 80 eV.

Microfluidic Resistive Pulse Sensing (MRPS) (Cimorelli et al., 2021): measurements of isolated extracellular vesicle size and concentration were performed using the nCS1™ instrument (Spectradyne LLC, United States), which is based on the microfluidic resistive pulse sensing technology. EVs isolated from 250 polyps and resuspended in Hydra medium were diluted in equal volume with Tween/PBS buffer in order to meet the instrument requirements regarding medium conductivity. EV samples were analysed with TS-400 and TS-300 cartridges (measurement range 60–400 nm and 50–300 nm, respectively).

Dynamic light scattering (DLS) was performed with a Malvern Zetasizer using EVs diluted in PBS at pH 7.4. Three runs were made at 25°C and at 90°scattering angle.

Protein concentration was determined using a colorimetric assay based on the Bradford Assay (ThermoFisher) and samples were diluted with PBS to obtain equal protein concentration (EV and whole Hydra homogenates). Samples were mixed with 4x Laemmli Sample Buffer containing 2-mercaptoethanol (Biorad) and boiled for 5 min at 95°C. Proteins were resolved by SDS-PAGE (mini protean TGX precast any KDa gel (Biorad)), blotted onto PVDF membranes (Amersham) and blocked o/n at 4°C in 5% non-fat dry milk in PBS-T (0.5% Tween-20). Membranes were then incubated with the primary antibody for 2 h at room temperature, washed with PBS-T 0.1%, and incubated with goat anti-rabbit IgG Antibody, (H + L) HRP-conjugated secondary antibody (1:4,000, BioRad) for 1 h at room temperature. Finally, protein bands were detected using an enhanced chemiluminescence reagent (Clarity ECL Substrate (Biorad)) and imaged with ChemiDoc XRS System (BioRad) using Image Lab software (BioRad). Antibodies: rabbit anti-Hydra Annexin B12 1:1,000 (kindly gifted by Dr. M. Isas, Keck School of Medicine of USC Los Angeles, CA), mouse anti-β-Actin−Peroxidase (1:40,000, Sigma).

EVs obtained as described before were labelled with PKH67 Green Fluorescent Cell Linker (Sigma Aldrich) following manufacturer’s protocol, with minor modifications. Briefly, EVs obtained from 500 polyps in HM were added to 1 ml of diluent C. As a negative control, 1 ml of diluent C was mixed with the same volume of HM. Afterwards, 6 μL of PKH67 dye was added and mixed for 30 sec by gently pipetting. After 5 min incubation at room temperature, 2 ml of 10% BSA was added to quench the unbound dye. EVs were washed and isolated by ultracentrifugation as described before. Hydra were incubated with PKH67- labelled EVs and with the negative control up to 5 h, washed with HM and imaged under a fluorescence microscope (Zeiss).

All the experiments were carried out using adult polyps starved for 24 h. For the head regeneration experiments, batches of 25 budless polyps were bisected at 80% body length (sub-hypostomal cut) or 50% body length (mid-gastric cut) and left regenerating for up to 72 h in presence of EVs isolated from 250 polyps (100 µL) in a total volume of 300 µL. The regeneration process was monitored at different time points using an optical microscope and polyps were grouped into four stages according to their tentacle morphogenetic features.

For the foot regeneration experiments, groups of 15 animals were bisected at 20% body length and transferred in HM containing EVs isolated from 250 polyps (100 µL) in a total volume of 300 µL. After 24 h, polyps were relaxed in 2% urethane in HM for 2 min and pre-fixed by adding 4% PFA/HM for 15 min. After removing the pre-fixation medium, animals were fixed for 1–3 h in 4% PFA/HM at 18°C and washed three times in PBS to remove any residual fixative. For the peroxidase assay, 2 ml of DAB solution was prepared according to the manufacturer’s instructions (Sigma-Aldrich tablets) and used for 15 min on the fixed animals. In order to stop the colorimetric reaction, the polyps were washed five times with ddH₂O and then mounted with PBS/Glycerol 1:1 on microscope slides for imaging.

Stained animals were observed by an inverted microscope (Axiovert 100, Zeiss) and images of each foot were taken with a digital camera (Olympus, DP 70) under the same conditions of acquisition (light and exposure time). The colorimetric analysis was performed using the HSB format on the acquired images in the ImageJ software version 1.53e. Three staining categories were arbitrarily set according to DAB positive area extension as low (0–9,000 μm²), medium (9,000–19,000 μm²) and high (19,000–28,000 μm²).

Prior to total RNA extraction, EVs isolated by ultracentrifugation were enzymatically treated to eliminate the free RNA derived from the ribonucleoprotein complexes that could be present in the conditioned medium. Isolated vesicles were treated with 1 mg/ml of proteinase K for 30 min at 37°C. Afterward, the samples were incubated 10 min at 65°C to inactivate the proteinase K and then treated with 10 µg/ml of RNAse A for 15 min at 37°C.

After treating polyps for 24 and 48 h with EVs isolated from 750 Hydra, total RNA was extracted from animals using Trizol or from EVs using Total RNA Extraction Kit (Norgen Biotek, Corp.) following manufacturer’s instructions and its concentration was determined using a SmartSpec plus spectrophotometer (Biorad). The first-strand cDNA was synthesized by High Capacity cDNA Reverse Transcription Kit (Applied Biosystem). qRT-PCR was performed using a Fast SYBR^TM Green Master Mix with premixed ROX (Applied Biosystems), in a StepOne Real-Time PCR System (Applied Biosystem) under the following fast cycling steps: initial denaturation for 20 s at 95°C, followed by 40 cycles at 95°C for 3 s, 60°C for 30 s. In addition, melting curves (20 min, from 59° to 90°C) were generated to check any undesired amplification products. EF1α was used as internal control.

In order to validate RNA-seq analysis, a presence/absence experiment of selected targets was performed using an EV sample isolated from 2000 polyps. Specific primers of Hydra genes Wnt3a, β-catenin, actin and Hsp70 were designed using the Primer3 software (http://frodo.wi.mit.edu/primer3/) and are listed in Supplementary Table S3, together with the corresponding GenBank accession numbers. Due to the qualitative nature of the presence/absence analysis, no internal control was used for the amplification reaction. The presence of such transcripts in the EV sample was confirmed by their cDNA amplification compared to negative controls, where no amplification occurred. Additional analysis by generation of melting curves (from 60 to 95, 0.5°C increment) confirmed sequence specific amplification in EVs and the absence of amplification in No Template Control (NTC). Mean C_t values for each gene and its NTC are shown in Supplementary Table S2

EVs were collected at different times and stored at −80°C. Total RNA extraction was performed on EV samples collected from 10,000 polyps overall and in two replicates using “Total RNA Purification Kit” (Norgen Biotek Corp.) and following the manufacturer’s supplementary protocol for exosomal RNA purification, with one modification: the optional DNA removal step was carried out using RQ1 RNase-Free DNAse (Promega) at 0.1 U/mL. The two samples were sent to the Sequentia Biotech S.L. (Barcelona, Spain) company for quality check, cDNA library preparation using “Ovation® SoLo RNA-Seq Systems” (NuGEN Technologies, Inc.), and sequencing via a paired-end chemistry on an Illumina platform. After quality inspection of the produced reads with fastqc (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore), reads were cleaned to remove the first five bases of read1, the Illumina adapter sequence “AGATCGGAAGAGC” and the bases with quality below 20 from the 3′ end using the program TrimGalore! (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore); the cleaned reads shorter than 35 bp were removed from further analysis (9.3% of the produced read pairs). Trimmed reads were then aligned using the program STAR (version 2.7.1a)⁷⁸ against the NCBI Hydra vulgaris reference genome (assembly Hydra_RP_1.0), while for gene expression quantification, trimmed reads were also aligned on the corresponding transcript sequences GCF_000004095.1_Hydra_RP_1.0_rna.fna.gz using the program kallisto (Dobin et al., 2013; Bray et al., 2016). Kallisto was also used to quantify genes using the reference transcriptome used in Siebert et al. (2019) in order to directly map the genes with the cell types where they were identified.

To perform the MALDI-TOF/TOF EV samples were first run into a SDS-Page gel. EV total proteins were quantified using a Bradford assay, 4x Laemmli Sample Buffer with 2-mercaptoethanol (Biorad) was added, samples were heated to 95°C for 5 min and they were run in a mini protean TGX precast 7.5% gel (Biorad). The gel was silver stained following standard procedures.

In-gel digestion- Protein bands were manually excised with a cutter, and in-gel digested with an automatic digestor (Intavis, Bioanalytical Instruments, Cologne, Germany). Briefly, spots were washed with water, ammonium bicarbonate (100 mM), and acetonitrile. Next, samples were reduced by incubation with DTT (10 mM) at 60°C for 45 min and alkylated by incubation with iodoacetamide (50 mM) at room temperature for 30 min. Finally, proteins were digested with trypsin overnight at 37°C (5 ng/μL, Trypsin Gold, Promega, WI, United States). Digestion was stopped by adding 0.5% TFA (trifluoroacetic acid), and tryptic peptides were extracted sequentially with increasing concentrations of acetonitrile in water. Peptides were concentrated and desalted by passing it through ZipTip C18 tips (Millipore) following the manufacturer’s instructions and eluted with 50%ACN 0.1%TFA.

Peptides were spotted into an Opti-Tof 384 well insert plate (Sciex) with a saturated solution of alpha-cyano-4-hydroxycinnamic acid (CHCA) prepared in 50% ACN/0.1% TFA. Proteins were identified using a 4800plus MALDI-TOF/TOF (Sciex) in the reflector mode with accelerating voltage of 20 kV, mass range of 800–4000 Da and 1,000 shots/spectrum. MS/MS spectra were acquired automatically on the 20 most intense precursors. Spectra were calibrated externally using a standard protein mixture (4,700 Calmix, ABSciex).

Proteins were identified with the search engine Mascot (Matrix Science, London, United Kingdom) against the NCBI prot database of Hydra (29,584 sequences). Search parameters used were: missed cleavage 1, fixed modifications carbamidomethyl (cysteines), variable modifications oxidation (methionine) and peptide and fragment mass tolerance 0.2 and 0.3 Da, respectively. Proteins with a score above 23 were considered a positive hit.

In solution digestion- Samples were evaporated and resuspended in 10 μL of denaturing buffer (6 M urea, 100 mM Tris buffer pH 7.8). Next, cysteines were reduced with 1.5 μL DTT (200 mM) for 30 min at 37°C and alkylated with 6 μL of iodoacetamide (200 mM) for 30 min in the dark. Unreacted iodoacetamide was consumed adding 6 μL of the reducing agent (200 mM DTT) for 30 min at room temperature. Samples were diluted with 50 mM ammonium bicarbonate to a final concentration less than 1 M of urea. Digestion was carried out overnight with the enzyme trypsin (Gold Trypsin, Promega) at 37°C and a 1:10 ratio (enzyme/protein). Reaction was stopped adding concentrated formic acid (Sigma). Samples were evaporated, resuspended in 2% acetonitrile, 0.1% formic acid and filtered through 0.45 μm filters. Protein identification was performed on a nano-LC system (Tempo MDLC, Eksigent, Dublin, CA, United States) coupled to a hybrid triple quadrupole/linear ion trap mass spectrometer (4000 QTRAP, Sciex). After precolumn desalting, tryptic digests (1 μg) were separated on a C18 column (Acclaim PepMap100, 75 μm id, 15 cm, 3 μm particle size, Thermo Scientific, United States) at a flow rate of 300 nL/min and a 120 min linear gradient from 5 to 35% ACN in 0.1% formic acid. The mass spectrometer was interfaced with a nanospray source equipped with uncoated fused silica emitter tip (20 μm inner diameter, 10 μm tip, NewObjective, Woburn, MA) and was operated in the positive ion mode. MS source parameters were as follows: capillary voltage 2800 V, source temperature 150°C, declustering potential (DP) 85 V, curtain and ion source gas (nitrogen) 20 psi, and collision gas (nitrogen) high. Analyses were performed using an information dependent acquisition (IDA) method with the following steps: single enhanced mass spectra (EMS, 400–1,400 m/z) from which the eight most intense peaks were subjected to an enhanced product ion [EPI (MS/MS)] scan. Proteins were identified with the search engine Mascot (v2.3 MatrixScience, United Kingdom) using NCBIprot database of Hydra (29,584 sequences). Search parameters used were: missed cleavages 1, fixed modifications carbamidomethyl (cysteines), variable modification oxidation (methionine) and peptide and fragment mass tolerance 0.5 and 0.3 Da, respectively. Only proteins with an overall Mascot score higher than 32 and at least two matching spectra were considered in the final list.

EVs were isolated from the medium of starved polyps by differential centrifugation, following a well-established protocol (C Thery et al., 2006), slightly modified. Pelleted EVs suspended in Phosphate buffered saline (PBS) where then characterized by Transmission Electron Microscopy (TEM). Figure 1C shows the round shape of isolated EV and the presence of double membrane around the electron dense vesicle (Supplementary Figure S1). By microfluidic resistive pulse sensing (MRPS) a yield of 5 × 10¹⁰ particles/mL was estimated from a typical preparation (250 polyps into 10 ml of medium), while the distribution of the particle size showed two peaks at 66 and 70.5 nm (Figure 1D and Supplementary Figure S1). A good correlation with these measurements was obtained through other approaches, such as Dynamic Light Scattering (DLS) analysis and TEM (Supplementary Figure S1), both confirming the average size of 66 nm, and suggesting an enrichment in exosome-like vesicles. The presence of EVs on Hydra ectodermal cell surface was previously supposed during an ultrastructural study describing several inward and outward mechanisms negotiating the entrance, trafficking, and clearance of gold nanoparticles in animal tissues (Marchesano et al., 2013). To confirm this evidence, Hydra polyps were pulsed (24 h) with positively charged AuNPs, synthesised as previously described (Marchesano et al., 2013) and 48 h later processed for TEM (Supplementary Figure S2). Semi-thin slices showed the presence of round nanovesicles decorated with AuNPs on the external membrane of ectodermal cells, laying into the glycocalix (Figures 1E,F), confirming in vivo assembly of AuNP-EV structures as part of the secretory route (Marchesano et al., 2013). This assembly demonstrates a strong interaction between EV membranes and positively charged AuNPs, likely due to electrostatic interaction.

The ability of EVs to be internalized into Hydra tissues was investigated by labelling freshly isolated EVs with PKH67 and incubating living polyps up to 5 h. Distinct green fluorescent signals were found on the body column and tentacles 2 h post incubation and in single cells obtained by maceration of treated polyps, indicating an efficient uptake of EVs into Hydra cells (Figures 1G–I; Supplementary Figure S3). These signals became hardly detectable after 5 h, possibly suggesting the processing of the EVs into the recipient cells.

Initially, a MALDI-TOF/TOF analysis performed on EV protein bands excised from an SDS-PAGE gel identified Annexin B12 and actin, whose presence was confirmed by immunoblot analyses (Supplementary Figure S4). Then a large scale EV isolation was achieved from 8,000 Hydra, and protein extracts were used for LC-ESI-MS/MS analysis and protein identification. The analysis from two biological replicates identified 52 proteins with a Mascot score >32 (Supplementary Table S1). Table 1 shows selected unique Hydra proteins grouped for function. Remarkably, 29 proteins were present also in the Exocarta protein database (http://www.exocarta.org/), suggesting the exosome–like nature of the isolated EVs. The exosome-specific proteins include two annexins, a synthenin1-like protein involved in the biogenesis of exosomes, a programmed cell death 6-interacting protein involved in the MVB formation and the CD151 antigen belonging to the Tetraspanin family. Among others, chaperones (Heat Shock Proteins), cytoskeletal proteins (Actin, Tubulin), extracellular matrix constituents (Fibrillin 1, Fibrillin 2), and important components of cell-cell adhesion (Protocadherin fat 4) and vesicle biogenesis (Synthenin, Programmed cell death 6-interacting protein, Polyubiquitin-B) were found. Altogether the variety of putative proteins identified in the Hydra EVs, either structural or involved in cell-cell communication and signal transduction, suggests multiple functional roles played by EVs in adult Hydra, and their correlation with known exosome proteomes.

A comprehensive transcriptomic analysis was performed on total RNA from purified EVs (from two biological replicates) generating two cDNA libraries and conducting pair-end next-generation sequencings. The sequenced reads were first aligned against the NCBI Hydra vulgaris reference genome (assembly Hydra_RP_1.0), where a percentage of 32 and 22% could be mapped. For gene expression quantification, the reads were also directly aligned using the program kallisto on the Hydra_RP_1.0 transcript sequences, where the total mapped reads on transcripts were 50 and 45% of the reads mapped on the genomic sequence. The presence of a subset of transcripts was confirmed in EV-derived RNA by using reverse transcription polymerase chain reaction PCR (RT-PCR). The selection was based on manual screening of genes already identified in EVs and exosomes (hsp70 and actin) or belonging to the Wnt/β−catenin signalling pathway (β−catenin, and Wnt3). We then used the minimum expression level in TPM (Transcripts Per Kilobase Million) of the validated genes to define an empirical detection threshold (0.4 TPM in at least one library). Using this definition, we selected 6,100 detected genes, among which 5,499 were protein-coding, 525 long non-coding and 31 tRNA genes. The detected genes were functionally annotated with the known associated Gene Ontology (GO) terms using OrthoDB to highlight the most important biological processes encoded by EV mRNAs (Kriventseva et al., 2019). The full list of the GO categorized genes is provided in Supplementary file 1. Interestingly, this analysis revealed that the most abundant transcripts within the Molecular Functions GO category are related to catalytic activity (alcohol dehydrogenase, sulfite oxidase), proteases and proteinase inhibitors (cathepsin, several members of the astacin family, antistasin), hydrolase and transferase activity. Within the Cellular Component GO category the genes encoding for intracellular and cytoplasmic anatomical structures were highly abundant, such as actin, cophylin, α-tubulin chain, β tubulin chain, talin-2, secreted glycoproteins (mucin 2-like, mucin 5AC-like), and LAMP, a well-known exosome marker. The transcripts categorized into the Biological processes GO class included the classes of protein metabolism (tRNA synthase, eIF3, eIF4, eEF1α, tRNAs), signal transduction pathways (G-protein coupled receptors, Rho family, guanidine nuclear exchange factors, Wnt3, gremlin) and response to stress (Hsp70, Hsp90, Superoxide dismutase, Glutamate dehydrogenase, thioredoxin-like).

Beside transcripts identification, lncRNA species were detected from the RNA-seq analysis. In one case the lncRNA was mapped adjacent to the mucin5AC gene locus, detected both at mRNA and protein levels, suggesting the possibility that EVs may shuttle all molecular components for immediate and late availability of key information in target cells. The association of lncRNA with exosomes was first demonstrated by the analysis of human plasma-derived exosomes by RNA-seq (Huang et al., 2013), and has been recently functionally characterized and linked to disease, including a list of cancers, where they act via epigenetic regulation of key target genes. Their presence in Hydra EV might suggest a regulatory function in the receiving cells.

In order to map the EV cellular source in the animal body we used a recently-published single-cell RNA-seq (scRNA-seq) atlas (Siebert et al., 2019) to match Hydra EV-associated genes with molecular signatures of specific cell clusters, which were grouped according to cell lineage (ectodermal, endodermal and interstitial cell lineage), or anatomical location (head, body column and foot) (Figure 2). An alignment rate of 12% was obtained and 4,997 transcripts were detected, of which 2,975 were found in at least one cell type in the scRNA annotated on Swiss-prot (TPM 0.2 in at least one library, Supplementary file 2). A large percentage of transcripts (48%) was clustered as specific of the interstitial cell lineage, while those derived from the ectodermal (12.7%) and endodermal cell lineage (9.45%) were less abundant. By subclustering the EV transcripts according to unique cell type expression, using the same cell categories and differentiation trajectories identified in the scRNA-seq (Siebert et al., 2019), a remarkable contribute of basal disk cells, battery cells, neurons and nematoblasts emerged (Figures 2B,C). A detailed analysis subclustering EV transcripts into cell types of each cell lineage is shown in Supplementary Figure S5. Overall, this indicates the mixture of precise molecular information from multiple cell types in Hydra EVs. The list of the five top genes in each cell cluster is provided in Supplementary Table S4 and the list of the 10 top genes identified matching the scRNAseq is shown in Supplementary Table S5. Interestingly, many genes belonging to the Wnt/β catenin signalling pathway were identified (Table 2), including Wnt3, β−catenin, naked cuticle (Petersen et al., 2015) together with some inhibitors such as dkk1/2/4 (Augustin et al., 2006; Guder et al., 2006), astacin (Ziegler et al., 2021), thrombospondin (Lommel et al., 2018), and sp5 (Vogg et al., 2019a). Remarkably, dkk1/2/4 and astacin were found among the most abundant transcripts specific of zymogen and granular mucous gland cell classes, strongly indicating their role in restricting the head organizer activity in the hypostome. By comparing the EV transcriptome to the proteome many annotated proteins (found also on exosome proteomes) were identified also at transcript level (actin, tubulin, annexinB12, hsp70), indicating the concomitant transfer of molecular information for immediate and programmed response in the target cells.

The finding in the EV transcriptome of both positive and negative regulators of the Wnt/β-catenin pathway (Figure 3A) prompted us to evaluate the EV bioactivity in receiving polyps. Wnt3 gene transcript levels were shown upregulated in polyps treated with EVs, both at 24 and 48 h post treatment (Figure 3B), suggesting for the first time the involvement of Hydra EVs in the exogenous activation of Wnt signalling.

Next, the bioactivity of the EV molecular cargo was evaluated in the context of the regeneration process. Freshly prepared EVs were added to polyps amputated at midgastric level (50% of body column length), and the head regeneration efficiency was evaluated by morphometric analysis. Stumps of progressive developmental stage (stages 0, 1, 2 and 3) were monitored at regular intervals, starting at 24 h post amputation (p.a.) (Figure 4A). At each time point significant differences were observed in the distribution of developmental stages, with the EV-treated stumps always presenting more advanced stages compared to the controls (Figure 4B). Similar results were also obtained when the amputation was sub-hypostomal (80% of upper body length) (Supplementary Figure S6). These data indicate a clear enhancement effect played by EVs in the regeneration of the head, suggesting that the EV cargo may act as an exogenous source of molecular signals. The impact of EVs on the ability of the lower body cells to regenerate a new foot was evaluated by bisecting whole polyps at 20% of the body length (Figure 4C). The absence of distinctive morphological changes that characterize this process led us to use biochemical assays to monitor the differentiation of foot specific cells, based on the detection of peroxidase activity. This enzymatic activity is used as specific marker of basal disk cell differentiation, producing a strong signal at 36–48 h p.a. (Hoffmeister and Schaller, 1985). While control regenerates showed a clear peroxidase-positive basal disk at 36 h p.a., there was only a faint staining in the basal disk region at this time, indicating an inhibitory effect played by EVs on foot regeneration process. The observed modulation of the regeneration confirms the presence of important messages into the EVs, as detected by the transcriptome analysis, and suggest a possible regulatory role in Hydra biology.